Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Aug 9.
Published in final edited form as: J Am Chem Soc. 2025 Dec 19;148(1):1926–1933. doi: 10.1021/jacs.5c19758

What Determines the Directionality of Kinesins? Studying the Anomalous Directionality of Nonclaret Disjunctional Motor Mutants

Ritaban Halder 1, Linfeng Hu 2, Arieh Warshel 3,*
PMCID: PMC13452579  NIHMSID: NIHMS2186492  PMID: 41414828

Abstract

NCD (nonclaret disjunctional) motors are a class of kinesin family proteins that are known to walk toward the minus end along the microtubule. Understanding the origin of such unique directionality of NCD motors and related kinesins has been a problem of fundamental importance. Experimental studies have indicated that specific NCD motor mutants switch their directionality. A type of NCD mutant (ran12) walks in the opposite direction (plus end) compared to the wild type (WT), whereas another mutant (N340K) of NCD shows diffusive bidirectionality. This work establishes that the directionality of these NCD motors is regulated by the barrier of the rate-determining ADP release step. We explore the molecular origin of the anomalous directionality of the mutated NCD motors compared with the wild type and determine the change in the barriers of ADP release for forward and backward motions. Our investigation is performed by combining different simulation techniques such as coarse-grained (CG) modeling, all-atom steered molecular dynamics (SMD) simulation, free energy estimation, as well as structural and stability analysis, and the maximum entropy (MaxEnt) method. We conclude that the differences in the directionality of different types of NCD motors are controlled by the free energy barrier for the ADP release step. Finally, we show that the statistical energy deduced from homologous sequences by the maximum entropy (MaxEnt) principle is significantly correlated with the stability and experimentally measured velocities of the wild type and several NCD mutants, and established that the stability loss of specific NCD mutants contributes to their reduced velocities and anomalous directionality on the microtubules. This provides a rare case where the origin of the AI prediction is clearly related to a physical-based picture.

Graphical Abstract

graphic file with name nihms-2186492-f0008.jpg

1. INTRODUCTION

NCD is a class of kinesin motor proteins that are involved in several important physiological processes, such as chromosome segregation during cell division and meiosis, proper spindle formation, and organization.1–4 The term “claret” in nonclaret disjunctional (NCD) refers to a specific type of genetic mutation in the fruit fly (Drosophila melanogaster), originally named after the wine-red eye color (claret) observed in the mutant flies. It belongs to the Kinesin-14 protein family. Kinesins are a class of motor proteins that are essential for the active transportation of cellular materials. Biomolecular migrations such as neuronal transport are one of the fascinating cellular phenomena that are often facilitated through the active transport mechanism by kinesin motors.5 Diffusion along neurons that are up to 1 m long (e.g., ischiadicus nerve or sciatic nerve) would take forever, and therefore, evolution invented active transport by kinesins6 to make such processes feasible and efficient. Unlike most of the kinesin family proteins that are known to move toward the plus end of microtubules, NCD moves toward the minus end of the microtubule track.7 They utilize ATP hydrolysis to generate force and movement along microtubules. It functions as a dimer and mainly consists of three domains, i.e., the motor head domain (consists of two head domains: front head and rear head), the neck linker region, which forms a coiled-coil stalk, and the tail domain, which is a basic proline-rich segment. NCD mutations often lead to the development of chromosomal aneuploidies.8 Therefore, understanding the functional aspects of NCD and the effect of its mutation is crucial to develop potential strategies to combat these abnormalities. The directionality of the movement of a motor protein is a key factor in its function and activity. The cytoskeletal motor protein dynein moves toward the minus end of the cell along microtubules,9,10 whereas the myosin family proteins move toward the plus end of the cell;11 the only exception is myosin VI, which has a minus-end directionality.12 It is important to note that although the functional activity of kinesin, dynein, and myosin motor proteins could differ, all of them show some key similarities in ATP hydrolysis and subsequent release of the hydrolyzed products (ADP and Pi) govern the cycle of their processes, which occur along either actin or microtubule tracks.13

Like many other ATPase motor proteins, kinesin motors generally move through a cycle of processes, as described schematically in Figure 1. The cycle is initiated by the attachment of the double-headed motor to the microtubule track, where one of the motor heads binds to ADP (Figure 1, step I). Then, the ATP molecule binds to a motor head domain (Figure 1, step II). Next, there is a conformational change in the motor head domain leads to a switch in the motor head to a lever-up conformation (Figure 1, step III). Then, the ATP hydrolysis occurs (Figure 1, step IV). Next, the trailing head moves forward (Figure 1, step V), which leads to the release of the bound phosphate and rebinding to the microtubule (Figure 1, step VI). Finally, the ADP release step (Figure 1, step VII) takes place, and the motor steps forward and carries cellular materials from one end to the other end inside its cellular destination. The cycle continues until the motor protein fulfills its purpose of cellular migration. It is important to note that in the case of NCD motors, ADP release is the slowest step throughout the whole cycle.14,15 The overall cycle generally controls the direction and average speed of the NCD motion. It is important to note in this context that the relationship between the energetics and dynamics of a motor protein (myosin) and its directionality has been explored extensively in our early studies16–18 and has been established that directionality is controlled by rate-determining barrier for forward and backward motions, i.e., it is the energetics rather than some unspecified dynamic effects that control the directionality.

Figure 1.

Figure 1.

Open in a new tab

A general schematic diagram of the mechanochemical cycle of the motor. The walking mechanism for a single step of the motor (dimer) along the microtubule filament toward the forward direction. The front head and rear head are marked by red and blue circles, respectively. In the case of the wild-type NCD motor, the forward movement corresponds to the minus-end directionality, whereas the backward movement corresponds to the plus-end directionality. The cycle initiates with the attachment of the double-headed motor to the microtubule track, where one of the motor heads is bound to ADP (step I). Then, the ATP molecule binds to the motor head domain (step II). Next, a conformational change in the motor head domain leads to a switch of the motor head to a lever-up conformation (step III). Then, the ATP hydrolysis occurs (step IV). Next, the trailing head moves forward (step V), which then leads to the release of bound phosphate and rebinds to the microtubule (step VI). Finally, the ADP release step (step VII) takes place, and the motor steps forward and carries cellular materials from one end to the other end inside its cellular destination. The cycle continues until the motor protein fulfills its purpose of cellular migration.

Many experimental and computational studies have been carried out to understand the function and activity of different kinesin family proteins, including the NCD motor.19–23 Sablin et al. used gliding motility assay experiments24 to show that specific NCD mutants have relatively slower velocities on the microtubule track and pointed out that a particular mutant among them (referred to as the ran12 mutant) exhibits reverse directionality compared to the wild-type NCD. Another study by Endow et al.,25 using low-density laser trap experiments, pointed out that a specific NCD mutant, N340K, shows diffusive bidirectionality on both ends along the microtubule without any particular directionality bias. Despite all of these and other experimental and computational efforts, the detailed mechanism and underlying molecular basis of the directionality switching of specific NCD mutants are still not properly understood. The directionality issue is also a central issue in understanding the action of myosin, where, despite relevant studies,16,18 the corresponding conclusions have not been widely accepted. An interesting study in this context by Astumian26 indicated that the directionality of a motor protein is governed by the relative heights of the energy barriers between different states. It was also pointed out that molecular recognition, which is the ability of a molecular machine to discriminate between the substrate and the product depending on the state of the machine, is more important for the determination of the inherent directionality and thermodynamics of chemo-mechanical coupling of the motor. An important study in this context27 indicates that power strokes can only determine the directionality of designed light-driven motors, but they do not play a significant role in any ground-state motor in the general context. It has been proposed that optically driven systems follow Einstein’s law of light absorption and emission, and directionality could be followed from an exergonic power stroke. On the other hand, ground-state ratchets are generally governed by the principle of microscopic reversibility, and they derive directionality from kinetic asymmetry, not from a power stroke.

Next, we focus on the experimental study where it was found that a 12-residue variant of the NCD motor (called the ran12 variant) shows plus-end directionality along the microtubule, which is the reverse of the general minus-end directionality of the wild-type NCD motor. From all-atom SMD analysis and free energy analysis, we find that the reverse directionality of the ran12 variant is due to its anomalous ADP release trend compared to that of the wild type. Our analyses showed that the wild-type NCD has a higher energy barrier of ADP release for the front end (prestroke state) compared to the rear end (poststroke state), which has also been observed experimentally. On the other hand, in the case of the ran12 variant, the energy barrier for ADP release is lower at the front end than at the rear end. Another reported NCD mutant is N340K, which shows bidirectionality (moving both plus and minus ends along the microtubules). It is demonstrated that in the case of N340K, the ADP release barrier from the front end and the rear end of that mutant did not show any specific preference, indicating no bias of ADP release from any specific end. This explains why the N340K variant of NCD exhibits bidirectionality without any specific directional bias. Further characterization indicates that specific interactions were missing in these NCD mutants, and such a loss of interaction is associated with their loss of stability as well as differences in the ADP release barrier. Finally, using evolutionary information comprising the maximum entropy approach,28 it is found that the experimentally measured velocities of different NCD mutants on microtubules, as well as the coarse-grained stabilities of those mutants, are well correlated with their maximum entropy values. We established that the stability loss of NCD mutants contributes to their reduced velocities as well as defective directionalities on microtubules, which in turn are modulated by the rate-limiting ADP release barrier in the NCD cycle.

As a complement to the above physical-based study, we also conducted an AI study in which we looked for a correlation between the velocity and Maximum Entropy (MaxEnt), as well as between the stability and MaxEnt. The results provide an example of a case in which the physical basis of an AI correlation is clearly established.

2. RESULTS AND DISCUSSION

2.1. All-Atom SMD Simulation of Wild-Type and Mutated NCD Motor.

We focused on the rate-limiting ADP dissociation energetics from the wild-type NCD motor and the ran12 and N340K variants of the NCD motor because these two mutants of NCD are known experimentally to exhibit anomalous directionality compared to the wild-type NCD. To understand how ADP dissociates from different states of the wild-type NCD motor protein and its ran12 and N340K variants, we first performed all-atom SMD simulations and monitored the time-dependent distance variation between the bound ADP molecule and these NCD motors. SMD provides a crude qualitative understanding of ligand dissociation from proteins.29,30 The details of the SMD simulation can be found in SI Appendix. The sequence difference between the wild-type NCD and the experimentally reported 12-residue mutants (ran12) is illustrated in Figure 2. As mentioned earlier, we considered both the poststroke and prestroke conformations of wild-type NCD motor and the variants to understand ADP dissociation from these NCD constructs. As shown in Figure 3, ADP is bound to NCD at lower distances, whereas ADP begins to dissociate from NCD at higher distances. We found that in the case of the wild-type NCD motor, ADP release is faster from the poststroke (rear head) conformation and slower from the prestroke (front head) conformation of the NCD (Figure 3a). This indicates that ADP dissociates more easily from the poststroke state of NCD than its prestroke conformation. Experimental findings have already established a faster ADP dissociation from the poststroke state of NCD motor compared to its prestroke state. Therefore, our SMD results support the experimental findings. Next, we found that in the case of the ran12 variant of NCD, the ADP dissociation trend is notably different compared to the wild-type NCD (Figure 3b). The Ran12 variant shows faster ADP release from the prestroke state compared to its poststroke state. On the other hand, the N340K variant of the NCD motor did not show any notable difference in ADP release either from its prestroke or poststroke state (Figure 3c). However, to acquire a detailed energetic insight into the ADP dissociation from these NCD constructs in different states, we performed a free energy analysis in the next sections.

Figure 2.

Figure 2.

Open in a new tab

Specific sequence of residues 335–346 of wild-type NCD motor protein and the ran12 variant. Each of the residues is marked with a different color code.

Figure 3.

Figure 3.

Open in a new tab

SMD simulation results of (a) wild-type NCD motor at prestroke and poststroke states and (b) ran12 variant of the NCD motor at prestroke and poststroke states, and (c) N340K variant of the NCD motor at prestroke and poststroke states. At lower distances, the ADP and NCD motors are bound to each other, whereas at higher distances, they become dissociated from each other. The motor directionalities of different NCD constructs on the microtubule track are indicated inside each plot. This indicates that the wild-type NCD motor shows ‘Minus-End’ directionality and the ran12 variant shows ‘Plus-End’ directionality. The N340K mutant shows diffusive ‘Bidirectional’ motility on microtubules.

2.2. Free Energy Analyses of Different NCD Motor Constructs.

SMD simulations indicated that ADP release from the poststroke state is more feasible than the prestroke state of the wild-type NCD motor protein. To obtain a more reliable quantitative estimation of the energetics of the rate-limiting ADP release steps of the NCD motor, we performed umbrella sampling-based free energy simulation by considering the distance between the NCD motor and ADP as an order parameter (Figure 4) in our free energy analysis. Umbrella sampling is a widely used method to study the conformational transition of proteins and for the estimation of the free energy of protein−ligand binding/unbinding processes.31–36 Details of the umbrella sampling simulation method can be found in the SI Appendix. Here, at a lower distance, the ADP is bound to the NCD, whereas at a higher distance, the ADP becomes unbound from the NCD. In the free energy analysis, the most stable minima correspond to the ADP-bound state of NCD. We found from the energy analysis that the ADP release from the poststroke state is easier than the prestroke state of wild-type NCD (Figure 4a). This finding is in accordance with our SMD simulation results, as well as with the experimental findings.15 Interestingly, the energetics of the ran12 variant of NCD follow a different trend compared to the wild type (Figure 4b). We found that ADP release from the prestroke state of ran12 follows a lower energy barrier as compared to its poststroke state. On the other hand, for the N340K variant of NCD, the ADP release barriers from the prestroke and poststroke states are not significantly different (Figure 4c). Structural inspection shows (Figure 5) that in the wild-type NCD motor, many favorable electrostatic and hydrophobic interactions, such as D344−K640, M343−L418, R335−D424, K336−D424, and N340−K640, were present. However, in the ran12 variant, such interactions were significantly diminished due to the mutation, and the N340K mutant also suffers a key electrostatic interaction loss between N340 and K640, which is present in the wild type. The stability loss of the mutants as compared to that of the wildtype NCD contributes to the differences in the ADP release barrier. Next, the coarse-grained stability analyses of wild-type NCD motor and several other experimentally reported mutants, as well as evolutionary analyses using the maximum entropy approach along with available experimental data of different NCD constructs, explain some interesting observations. It is worth mentioning that in our previous work,18 we observed a tight bond between ADP and Mg2+ in a related ATPase motor protein system (Myosin V and VI). This bond raises the possibility that Mg2+ may be released along with ADP. Therefore, we demonstrated the effect of Mg2+ on ADP release and performed free energy calculations for the ‘Mg-ADP release’ and ‘ADP release.’ The corresponding free energy profile indicated that the release of the MG-ADP system and ADP alone follows a qualitatively similar trend.

Figure 4.

Figure 4.

Open in a new tab

Umbrella sampling simulation results for (a) wild-type NCD motor protein at the poststroke (rear head) and prestroke (front head) conformations, and the same for (b) the ran12 variant of the NCD motor protein and (c) N340K variant of the NCD motor protein. Here, the free energy is plotted by considering the distance between the ADP molecule and the NCD motor as an order parameter. Error bars were calculated using bootstrapping. The motor directionalities of different NCD constructs on the microtubule track were mentioned inside each plot.

Figure 5.

Figure 5.

Open in a new tab

(a) Important interactions in wild-type NCD motors in the region of residues 335–346. (b) Snapshot of these interactions. Note that in the case of the ran12 variant, many interactions were missing because of a large pool of mutations. Also, the N340K mutation eliminates the N340–K640 interaction, which is present in the wild type.

2.3. Maximum Entropy Model Predicts Structural Stability and Experimental Directional Motility of Different NCD Variants.

Next, we used the maximum entropy (MaxEnt) model (Figure 6) to assess whether the evolutionary information captured by the MaxEnt method28,37,38 is predictive of the experimentally measurable biophysical properties of NCD motor variants. Therefore, we analyzed two key phenotypes: structural stability and directional motility. The details of the model can be found in our previous studies28,37 and in the SI Appendix.

Figure 6.

Figure 6.

Open in a new tab

Schematic representation of the use of the MaxEnt model for NCD motor protein dynamics. A pairwise MaxEnt model was constructed from the MSA, and each of the protein sequences (S) is associated with a statistical energy (EMaxEnt) that follows the Boltzmann distribution. We found that decreasing the statistical energy derived from the MaxEnt model significantly correlates with the stability of different variants of the NCD motor protein and their walking velocity on the microtubule track.

Before we consider the MaxEnt analysis, we should mention that the stability is fundamentally related to the motor protein velocity/motility because the motor’s structure and its conformational changes are directly tied to its ability to perform its mechanochemical cycle (as evident from Figure 1 of the cycle), converting chemical energy derived from ATP into directed movement at a particular direction on its track. A less stable protein may detach more frequently from the track, leading to shorter “run lengths on the track” and effectively decreasing the average velocity of the cargo transport. A decrease in stability, often caused by mutations, can disrupt this cycle, leading to reduced motility. An interesting study39 in a related context by Brendza et al. indicates that a mutation of a single amino acid (T291M) in the core motor domain of Drosophila Kinesin leads to a complete loss of motor function both in vitro and in vivo, which is caused by the destabilization that weakens ATP binding and uncouples the activity between the two motor heads, hence disrupting the coordinated walking mechanism by significantly reducing its motility.

With the above considerations in mind, we calculated the structural stability of the wild-type NCD and its different variants. This was done using a coarse-grained method, and the velocities and directionalities of these variants have already been reported in earlier motility assay experiments (Table S1). In Figure 7, we show that the statistical energy derived from the MaxEnt model, trained on the NCD motor’s homologous sequences, correlates with both the computed stability changes (ΔEstability) and the observed motor velocity directionality. In the analysis of EMaxEnt versus calculated stability (Figure 7A), we observed a strong positive correlation (Pearson r = 0.833), suggesting that sequences with lower statistical energies tend to be more structurally stable. These results indicate that the MaxEnt energy landscape, although purely statistical, can qualitatively reflect the thermodynamic trends among rationally designed variants. Notably, the NCD-ran12 mutant, which replaces 12 conserved residues with a randomized hydrophilic sequence, deviates evidently from the wild-type but still aligns with the overall trend. Its inclusion in the regression suggests that the model may be extrapolated to larger sequence perturbations in some cases, although such extrapolations should be interpreted with caution. When comparing EMaxEnt with directional motility (Figure 7B), we observed a moderate negative correlation among the minus-end-directed variants (Pearson’s r = 0.697). In this case, NCD-ran12 was excluded from the regression analysis due to its synthetic design and reversed motility. Unlike the other variants, which were generated through stepwise or structure-informed mutations, NCD-ran12 represents a discontinuous and artificial departure from the native sequence space. A key observation is that the ran12 variant exhibits reverse directionality relative to the WT, being the least stable and slowest along the microtubules. It also shows the highest MaxEnt energy and a reverse ADP release trend compared to WT. Together, these results suggest that EMaxEnt encodes evolutionarily relevant constraints that can be related to physical properties such as stability and directional behavior. The details about all the mutations, their corresponding motor directionality, velocity/motility percentage, stability, and MaxEnt values are depicted in Supplementary Dataset 1. A list of all the sequences of NCD motor proteins used in this study is provided in Supplementary Dataset 2. Although EMaxEnt is not a direct predictor of function, it offers a promising framework for probing evolutionary sequence landscapes and guiding the interpretation of both natural and engineered mutations.

Figure 7.

Figure 7.

Open in a new tab

Evolutionary statistical energy (EMaxEnt) correlates with the stability and motor directionality of NCD variants. (A) Correlation between the maximum entropy statistical energy (EMaxEnt) and the calculated change in CG free energy for the NCD variants. A significant positive correlation is observed (Pearson r = 0.833). The red star indicates NCD-ran12, a rationally randomized mutant replacing 12 conserved residues in the neck region with hydrophilic residues (ESGAKQGEKGES), which causes a drastic loss of stability and a reversal of motor direction. All variants, including NCD-ran12, are included in the linear regression to assess the extent to which E_MaxEnt captures the overall stability trends in the system. (B) Signed velocity percentage versus EMaxEnt. A negative correlation is observed among the minus-end-directed variants (Pearson’s r = 0.697), which were constructed through rational, structure-based mutations. NCD-ran12 is shown for reference but is excluded from the regression due to its unique and nongradual design. Unlike other variants, NCD-ran12 is not part of a natural or structurally inferred mutational trajectory, and its reversal of motility highlights its distinct behavior in the sequence-function landscape.

3. CONCLUDING REMARKS

This work aims to determine the origin of the anomalous directionality of different types of NCD motor constructs. First, we systematically estimated the ADP release barrier of different NCD constructs in the forward and backward paths corresponding to the prestroke and poststroke states, respectively, using SMD analysis and all-atom free energy-based simulations. We found that SMD simulation results, as well as free energy-based estimation of ADP release barrier of the wild-type NCD motor, match well with the experimental trend, which also indicates that ADP release from the front head of the NCD motor possesses a higher energy barrier than that of the corresponding rear head.15 On the other hand, we explored the ADP release barriers of two different experimentally reported NCD mutants; one of them (ran12 construct) shows reverse (plus end) directionality compared to the wild type, whereas another construct (N340K) shows diffusive bidirectionality without any clear directionality bias. We found that ran12 exhibits the opposite trend of the ADP release barrier compared to the wild type, which explains that the reverse directionality of the ran12 construct is due to the different trend in the ADP release barrier compared to the wild type. The ADP release barrier of the N340K construct did not show any notable energy difference between the forward and backward paths, which explains why the experimentally observed diffusive bidirectionality of this construct is due to the lack of such energy-driven push for the ADP release barrier from any specific state. Our studies explored how the energy landscape of ADP release from different states of the NCD motor varied for the forward and backward motions. Therefore, the ADP release energetics from different states of the NCD motor protein are certainly related to the directionality of the NCD.

Here, we would like to point out that our earlier extensive free energy studies40,41 involving coarse-grained simulations not only provided a clear picture of the conversion of chemical energy to mechanical energy by motor proteins but also indicated that such key processes follow a fundamental principle of physics called ‘microscopic reversibility42 and do not involve any dynamical effects. We also found that in the case of the NCD mutants, significant destabilization due to the loss of electrostatic interactions leads to an alteration of the ADP release barrier. The trend of the ADP release barriers of different states of the wild-type NCD motor protein matches the experimentally reported trend, thereby establishing the robustness of our methods.

Finally, we explored the relationship between the evolutionary information on the NCD motor protein and the stability of different experimentally reported NCD mutants and their corresponding velocities on the microtubule track by correlating EMaxEnt obtained from their natural homologous sequences. We noticed that the decreased velocities of different NCD mutants on the microtubule track are related to their stabilities as well as to their evolutionary information, as evident from their correlation in this study. This is particularly important because such a correlation could inspire us to think about a way to efficiently design motor protein; i.e., we could consider a pragmatic and sensible “engineering approach” considering the correlation between EMaxEnt and the reported velocities of motor proteins and their corresponding stabilities. To be more precise, we can generate new motor protein mutants depending on the correlations, determine their EMaxEnt, and further screen them based on their rate-limiting ADP release barrier from different states and stability analysis, and then design experiments to obtain efficient motor proteins.

Thus, the methods used here could serve as pivotal methods for consensus motor protein design. The findings here seem to be the first reported study that provides a hint at how novel molecular motors could be designed using an evolutionary approach. A thorough examination of other motor proteins is required to obtain a generalized view of the implications of such a design principle. Such directed evolution-based rational designing of novel motor proteins could not only bolster our understanding of this field of research but also be important for efficient targeted drug delivery to specific locations inside the cell, as motor proteins serve as precursors for cargo delivery. In summary, MaxEnt is used as a correlator of the velocities/efficiencies of different NCD motor proteins, as well as their stabilities. Such a connection could open up new avenues in the fields of evolutionary biology as well as in motor protein engineering research.

It is interesting to comment on the general problem of AI. In other words, AI does not tell us the reason for the given correlation. However, in the present work, we have a remarkable example where a physical-based study provides the reason for the change in directionality and velocity, where the deduced correlation is subsequently reproduced by the MaxEnt treatment.

Supplementary Material

Supporting Information
Supplementary Dataset 1

Supplementary Dataset 1: Details about all mutations, their corresponding motor directionality, velocity/motility percentage, stability, and MaxEnt values (XLSX)

Supplementary Dataset 2

Supplementary Dataset 2: The list of all the sequences of NCD motor proteins used in this study (TXT)

Figure S1
Table S1

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.5c19758.

All-atom steered molecular dynamics methods, umbrella sampling free energy calculation methods, coarse-grained scheme for stability analysis, maximum entropy calculation methods, experimental velocities and directionalities of different NCD motor protein constructs (PDF)

ACKNOWLEDGMENTS

This work was supported by NIH Grant R35 GM122472 and NSF Grant MCB 1707167. We thank the University of Southern California High Performance Computing and Communication Center for computational resources. R.H. is thankful to Dr. Raphael Alhadeff and Dr. Zhen Tao Chu for many helpful discussions.

Footnotes

ASSOCIATED CONTENT

Complete contact information is available at: https://pubs.acs.org/10.1021/jacs.5c19758

The authors declare no competing financial interest.

Contributor Information

Ritaban Halder, Department of Chemistry, University of Southern California, Los Angeles, California 90089-1062, United States.

Linfeng Hu, Department of Chemistry, University of Southern California, Los Angeles, California 90089-1062, United States.

Arieh Warshel, Department of Chemistry, University of Southern California, Los Angeles, California 90089-1062, United States.

Data Availability Statement

The processed data generated in this study, such as the example of input files, force field parameters, coordinate files and structural data, topology and parameter files, postprocessing and pulling files, TMD and umbrella sampling data, energies, coarse-grained data, required scripts and codes, and sample simulation trajectories, were provided in the data repositories https://zenodo.org/records/17613829. Additional data regarding the MaxEnt codes and scripts can be accessed from the following data repositories https://github.com/EvoCatalysis/MaxEnt-Pytorch.

REFERENCES

  • (1).Endow SA; Komma DJ Centrosome and Spindle Function of the Drosophila Ncd Microtubule Motor Visualized in Live Embryos Using Ncd-GFP Fusion Proteins. J. Cell Sci. 1996, 109 (10), 2429–2442. [DOI] [PubMed] [Google Scholar]
  • (2).Chandra R; Salmon ED; Erickson HP; Lockhart A; Endow SA Structural and Functional Domains of the Drosophila Ncd Microtubule Motor Protein. J. Biol. Chem. 1993, 268 (12), 9005–9013. [PubMed] [Google Scholar]
  • (3).Fink G; Hajdo L; Skowronek KJ; Reuther C; Kasprzak AA; Diez S The Mitotic Kinesin-14 Ncd Drives Directional Microtubule−Microtubule Sliding. Nat. Cell Biol. 2009, 11 (6), 717–723. [DOI] [PubMed] [Google Scholar]
  • (4).Hallen MA; Liang Z-Y; Endow SA Ncd Motor Binding and Transport in the Spindle. J. Cell Sci. 2008, 121 (Pt 22), 3834–3841. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (5).Hayashi K; Sasaki K Number of Kinesins Engaged in Axonal Cargo Transport: A Novel Biomarker for Neurological Disorders. Neurosci. Res. 2023, 197, 25–30. [DOI] [PubMed] [Google Scholar]
  • (6).Encalada SE; Szpankowski L; Xia C; Goldstein LSB Stable Kinesin and Dynein Assemblies Drive the Axonal Transport of Mammalian Prion Protein Vesicles. Cell 2011, 144 (4), 551–565. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (7).Furuta K; Toyoshima YY Minus-End-Directed Motor Ncd Exhibits Processive Movement That Is Enhanced by Microtubule Bundling In Vitro. Curr. Biol. 2008, 18 (2), 152–157. [DOI] [PubMed] [Google Scholar]
  • (8).Hatsumi M; Endow SA Mutants of the Microtubule Motor Protein, Nonclaret Disjunctional, Affect Spindle Structure and Chromosome Movement in Meiosis and Mitosis. J. Cell Sci. 1992, 101 (3), 547–559. [DOI] [PubMed] [Google Scholar]
  • (9).Roberts AJ; Kon T; Knight PJ; Sutoh K; Burgess SA Functions and Mechanics of Dynein Motor Proteins. Nat. Rev. Mol. Cell Biol. 2013, 14 (11), 713–726. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (10).Golcuk M; Yilmaz SZ; Yildiz A; Gur M The Mechanism and Energetics of the Dynein Priming Stroke. Structure 2024, 32 (5), 603–610. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (11).Cappello G; Pierobon P; Symonds C; Busoni L; Christof J; Gebhardt M; Rief M; Prost J Myosin V Stepping Mechanism. Proc. Natl. Acad. Sci. U.S.A. 2007, 104 (39), 15328–15333. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (12).Sweeney HL; Park H; Zong AB; Yang Z; Selvin PR; Rosenfeld SS How Myosin VI Coordinates Its Heads during Processive Movement. EMBO J. 2007, 26 (11), 2682–2692. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (13).Sweeney HL; Holzbaur ELF Motor Proteins. Cold Spring Harb. Perspect. Biol. 2018, 10 (5), No. a021931. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (14).Lockhart A; Cross RA; McKillop DF ADP Release Is the Rate-limiting Step of the MT Activated ATPase of Non-claret Disjunctional and Kinesin. FEBS Lett. 1995, 368 (3), 531–535, DOI: 10.1016/0014-5793(95)00723-M. [DOI] [PubMed] [Google Scholar]
  • (15).Cochran JC Kinesin Motor Enzymology: Chemistry, Structure, and Physics of Nanoscale Molecular Machines. Biophys Rev. 2015, 7 (3), 269–299. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (16).Alhadeff R; Warshel A Reexamining the Origin of the Directionality of Myosin V. Proc. Natl. Acad. Sci. U.S.A. 2017, 114 (39), 10426–10431. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (17).Mukherjee S; Alhadeff R; Warshel A Simulating the Dynamics of the Mechanochemical Cycle of Myosin-V. Proc. Natl. Acad. Sci. U.S.A. 2017, 114 (9), 2259–2264. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (18).Halder R; Chu ZT; Ti R; Zhu L; Warshel A On the Control of Directionality of Myosin. J. Am. Chem. Soc. 2024, 146, 28360–28374, DOI: 10.1021/jacs.4c09528. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (19).Deguchi T; Iwanski MK; Schentarra E-M; Heidebrecht C; Schmidt L; Heck J; Weihs T; Schnorrenberg S; Hoess P; Liu S; Chevyreva V; Noh K-M; Kapitein LC; Ries J Direct Observation of Motor Protein Stepping in Living Cells Using MINFLUX. Science 2023, 379 (6636), 1010–1015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (20).Dietrich KA; Sindelar CV; Brewer PD; Downing KH; Cremo CR; Rice SE The Kinesin-1 Motor Protein Is Regulated by a Direct Interaction of Its Head and Tail. Proc. Natl. Acad. Sci. U.S.A. 2008, 105 (26), 8938–8943. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (21).Nitzsche B; Dudek E; Hajdo L; Kasprzak AA; Vilfan A; Diez S Working Stroke of the Kinesin-14, Ncd, Comprises Two Substeps of Different Direction. Proc. Natl. Acad. Sci. U.S.A. 2016, 113 (43), E6582–E6589, DOI: 10.1073/pnas.1525313113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (22).Jana B; Hyeon C; Onuchic JN The Origin of Minus-End Directionality and Mechanochemistry of Ncd Motors. PLoS Comput. Biol. 2012, 8 (11), No. e1002783. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (23).Lakkaraju SK; Hwang W Hysteresis-Based Mechanism for the Directed Motility of the Ncd Motor. Biophys. J. 2011, 101 (5), 1105–1113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (24).Sablin EP; Case RB; Dai SC; Hart CL; Ruby A; Vale RD; Fletterick RJ Direction Determination in the Minus-End-Directed Kinesin Motor Ncd. Nature 1998, 395 (6704), 813–816. [DOI] [PubMed] [Google Scholar]
  • (25).Endow SA; Higuchi H A Mutant of the Motor Protein Kinesin That Moves in Both Directions on Microtubules. Nature 2000, 406 (6798), 913–916. [DOI] [PubMed] [Google Scholar]
  • (26).Astumian RD Irrelevance of the Power Stroke for the Directionality, Stopping Force, and Optimal Efficiency of Chemically Driven Molecular Machines. Biophys. J. 2015, 108 (2), 291–303. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (27).Astumian RD Interpreting Molecular Ratchets, Directionality, and Kinetic Asymmetry through the Work of Curie, Einstein, and Lewis. Commun. Phys. 2025, 8 (1), 365. [Google Scholar]
  • (28).Xie WJ; Asadi M; Warshel A Enhancing Computational Enzyme Design by a Maximum Entropy Strategy. Proc. Natl. Acad. Sci. U.S.A. 2022, 119 (7), No. e2122355119. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (29).Patel JS; Berteotti A; Ronsisvalle S; Rocchia W; Cavalli A Steered Molecular Dynamics Simulations for Studying Protein−Ligand Interaction in Cyclin-Dependent Kinase 5. J. Chem. Inf. Model. 2014, 54 (2), 470–480. [DOI] [PubMed] [Google Scholar]
  • (30).Skovstrup S; David L; Taboureau O; Jørgensen FS A Steered Molecular Dynamics Study of Binding and Translocation Processes in the GABA Transporter. PLoS One 2012, 7 (6), No. e39360. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (31).Govind Kumar V; Polasa A; Agrawal S; Kumar TKS; Moradi M Binding Affinity Estimation from Restrained Umbrella Sampling Simulations. Nat. Comput. Sci. 2023, 3 (1), 59–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (32).Tam NM; Nguyen TH; Ngan VT; Tung NT; Ngo ST Unbinding Ligands from SARS-CoV-2 Mpro via Umbrella Sampling Simulations. R. Soc. Open Sci. 2022, 9 (1), No. 211480. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (33).Halder R; Warshel A How Omecamtiv Modulates Myosin Motion. Biochemistry 2025, 64 (10), 2318–2331. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (34).Halder R; Warshel A Energetic and Structural Insights behind Calcium Induced Conformational Transition in Calmodulin. Proteins 2024, 92 (3), 384–394. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (35).Halder R; Manna RN; Chakraborty S; Jana B Modulation of the Conformational Dynamics of Apo-Adenylate Kinase through a π–Cation Interaction. J. Phys. Chem. B 2017, 121 (23), 5699–5708. [DOI] [PubMed] [Google Scholar]
  • (36).Halder R; Warshel A Understanding Hypertrophic Cardiomyopathy and Its Regulation by Myosin Drugs. Protein Sci. 2025, 34 (10), No. e70289. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (37).Xie WJ; Warshel A Natural Evolution Provides Strong Hints about Laboratory Evolution of Designer Enzymes. Proc. Natl. Acad. Sci. U.S.A. 2022, 119 (31), No. e2207904119. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (38).Gelfand N; Orel V; Cui W; Damborsky J; Li C; Prokop Z; Xie WJ; Warshel A. Biochemical and Computational Characterization of Haloalkane Dehalogenase Variants Designed by Generative AI: Accelerating the SN2 Step. J. Am. Chem. Soc. 2025, 147 (3), 2747–2755. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (39).Brendza KM; Sontag CA; Saxton WM; Gilbert SP A Kinesin Mutation That Uncouples Motor Domains and Desensitizes the Gamma-Phosphate Sensor. J. Biol. Chem. 2000, 275 (29), 22187–22195. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (40).Bai C; Asadi M; Warshel A The Catalytic Dwell in ATPases Is Not Crucial for Movement against Applied Torque. Nat. Chem. 2020, 12 (12), 1187–1192. [DOI] [PubMed] [Google Scholar]
  • (41).Mukherjee S; Bora RP; Warshel A Torque, Chemistry and Efficiency in Molecular Motors: A Study of the Rotary−Chemical Coupling in F 1 -ATPase. Q. Rev. Biophys. 2015, 48 (4), 395–403. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (42).Astumian RD; Mukherjee S; Warshel A The Physics and Physical Chemistry of Molecular Machines. ChemPhysChem 2016, 17 (12), 1719–1741. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting Information
NIHMS2186492-supplement-Supporting_Information.pdf (204.7KB, pdf)
Supplementary Dataset 1

Supplementary Dataset 1: Details about all mutations, their corresponding motor directionality, velocity/motility percentage, stability, and MaxEnt values (XLSX)

NIHMS2186492-supplement-Supplementary_Dataset_1.xlsx (10.9KB, xlsx)
Supplementary Dataset 2

Supplementary Dataset 2: The list of all the sequences of NCD motor proteins used in this study (TXT)

NIHMS2186492-supplement-Supplementary_Dataset_2.txt (3.8MB, txt)
Figure S1
NIHMS2186492-supplement-Figure_S1.png (106.5KB, png)
Table S1
NIHMS2186492-supplement-Table_S1.png (73.1KB, png)

Data Availability Statement

The processed data generated in this study, such as the example of input files, force field parameters, coordinate files and structural data, topology and parameter files, postprocessing and pulling files, TMD and umbrella sampling data, energies, coarse-grained data, required scripts and codes, and sample simulation trajectories, were provided in the data repositories https://zenodo.org/records/17613829. Additional data regarding the MaxEnt codes and scripts can be accessed from the following data repositories https://github.com/EvoCatalysis/MaxEnt-Pytorch.

RESOURCES